Introduction

Late gadolinium enhancement (LGE) is the current gold standard imaging technique for the assessment of myocardial scars¹. Although it provides very good contrast between healthy and scar tissue, scar patterns may be confused with blood signal, resulting in ambiguities in scar detection. A novel imaging technique, termed SPOT², combines bright- (BR) and black- (BL) blood imaging in a single sequence to obtain images that removes this ambiguity and with improved myocardial contrast and scar localization. However, many patients with cardiac implantable electronic device (CIED) do not undergo cardiac LGE imaging due to severe regional hyperintensity artefacts³ induced by the devices. Here we propose a technology to enable imaging of myocardial scars with unprecedent scar contrast in patients with CIED by combining joint bright- and black-blood SPOT imaging with wideband MRI³.

Methods

Acquisition: SPOT² is a single-shot ECG-triggered sequence that collects BR and BL images in an interleaved fashion (Fig.1A). In odd heartbeats, BL images (providing better scar detection) are generated by a non-selective adiabatic hyperbolic secant (HS) 180° inversion recovery (IR) pulse followed by a T2 preparation module (T2p). In even heartbeats, BR images (providing better scar localization) are collected by only playing out a T2p. The T2p (duration=26.6ms) consists of a 90° tip-down pulse, of two adiabatic HS refocusing pulses, and of a 90° tip-up⁵. A crusher gradient removes all residual transverse magnetization.
To eliminate the hyperintensity artefacts from the non-properly inverted myocardium signal in the presence of CIED, the IR bandwidth was increased from 0.8 to 9.2kHz (Fig.1BC) with a peak B1 amplitude of 30μT. The IR duration was set to 10.24ms. This wideband IR was implemented in the existing reference 2D phase-sensitive inversion recovery (PSIR)¹ and in 2D SPOT,² referred to as wideband IR SPOT. To further reduce CIED-related artefacts, the bandwidth of the two T2p refocusing pulses were also increased from 1.6 to 5kHz (Fig.1BD) with a peak B1 amplitude of 30μT. This method is referred to as wideband IR+T2 SPOT hereafter. To eliminate bSSFP-associated banding artefacts in the presence of CIED, a gradient echo readout was used for both PSIR and SPOT⁴.

Experiments and analysis: The proposed sequence was tested in the ISMRM-NIST phantom (QualibreMD) and in one patient implanted with a CIED. The following parameters were used for all sequences: FOV=300mm², resolution=1.4x1.4mm², slice thickness=8mm, TE/TR=2.7/4.8ms, flip angle=15°, readout bandwidth=751Hz/pixel, 5 signal averages. All experiments were performed on a 1.5T system (MAGNETOM Aera, Siemens). The phantom study was conducted to test the effect of wideband IR SPOT and wideband IR+T2 SPOT, with an inversion time TI=200ms. Wideband SPOT was then compared to wideband PSIR, with TI=450ms for PSIR. A CIED was taped on the side of the phantom (Fig.2A). The extent of the artefact was measured on coronal images. Percentage artefact reduction was evaluated as the percentage reduction in banding and hyperintensity artefacts of each wideband sequence compared to the sequence without wideband. Signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were calculated. CNR was calculated in two phantom vials with T1/T2 values characteristic of human myocardium (green tube in Fig.3A) and scar (pink tube in Fig.3A). Finally, reference PSIR and wideband IR+T2 SPOT were tested in a cardiac patient implanted with a pacemaker, 15min after injection of 0.2 mmol/kg gadolinium-contrast based agent.

Results

In the presence of CIED, SPOT and reference PSIR provide phantom images with banding, hyperintensity artefacts and signal voids (Fig.2AB). The extent of the signal void for all sequences corresponded to 7.8cm compared to ICD-free imaging without artifacts. The artefact extent with wideband PSIR decreased from 14.5 to 7.8cm (white arrows, Fig.2A), with hyperintensity artefacts entirely suppressed but remaining signal void (7.8 cm extent) (Fig.2A). CNR slightly improved with wideband PSIR whereas SNR slightly decreased (Fig.3A). The artefact extent in conventional SPOT BL and BR was of 14.5cm and 12cm, respectively. Wideband IR SPOT reduced the artefact extent to 11.7cm on BL images but with remaining hyperintensity artefacts (91% reduction). As expected, wideband IR had no effect on BR images. However, wideband IR+T2 SPOT reduced the artefact extent on both BL and BR images to 7.8cm, with suppressed hyperintensity and banding artefacts. CNR slightly decreased for wideband IR SPOT and wideband IR+T2 SPOT on both BL and BR images. SNR was similar for conventional, wideband IR and wideband IR+T2 on BL and BR images. Overall, wideband IR+T2 SPOT performed as good as wideband PSIR for the reduction of hyperintensity artefacts, but with a much higher CNR on BR and BL images and higher SNR on BR images.
In the patient, banding and hyperintensity artefacts are present when using conventional PSIR but were successfully suppressed with wideband IR+T2 SPOT (Fig.4).

Conclusion

Wideband IR+T2 SPOT can suppress hyperintensity and banding artefacts on both black- and bright-blood images compared with conventional PSIR and wideband IR SPOT, making it a promising technique for scar assessment in patients with CIED. In vivo clinical testing is now warranted.